The phone call that lands at Patientdesk.ai's system sounds unremarkable at first—a prospective patient wondering whether their insurance will cover a dental visit, hesitating before committing to an appointment. But the voice answering isn't human. And unlike most healthcare AI demos, this one is fielding real questions from real patients, live across more than 60 clinics scattered between the US, Australia, and the UK.
The Y Combinator-backed startup has built what it describes as the first AI voice agent capable of verifying insurance coverage during the call itself, not hours later via callback. It handles scheduling, collects payments, and books directly into a clinic's existing software—all before the patient hangs up. Whether that's genuinely a category first or just clever positioning is debatable, but the ambition is clear.
Three founders—Oncel Ozgul, Fikri San Koktas, and Emre Kaplaner—announced a $1 million pre-seed round in mid-February, led by Y Combinator with participation from E2VC. Their pitch strips away the usual healthcare AI abstractions: automate the front desk workflows that chew through staff hours and bleed revenue every time a practice misses a call.
The Real-Time Insurance Gamble
Most voice AI systems in healthcare punt insurance questions to callbacks or defer to staff. Patientdesk.ai's founders are betting the opposite—that solving coverage questions on the spot is what converts hesitant callers into booked appointments.
When a prospective patient rings, the AI doesn't just take a message. It checks insurance eligibility through a clearinghouse integration, quotes out-of-pocket costs, collects payment or a deposit if required, and slots the appointment directly into the practice's calendar. One call, theoretically no friction.
The system integrates with major dental platforms—Dentrix, OpenDental, Eaglesoft, Dentalicon—plus broader EHR systems like athenahealth, eClinicalWorks, and NextGen. The company's FAQ claims it can work with "any major PMS that has an online appointment page or API," though formal partnership announcements remain sparse. A LinkedIn post recently confirmed a tie-up with Core Practice, an Australian practice management vendor.
Beyond inbound scheduling, the platform runs outbound follow-ups and payment reminders around the clock. The company bundles this with paid lead generation through Meta ads, positioning itself as both receptionist and growth engine—a combination that blurs the line between operational tool and marketing spend.
The Numbers (With Caveats)
In their Y Combinator launch post, the team shared an anecdote: one clinic using the system generated $350,000 in revenue during December from Patientdesk-handled bookings. The company also claims it scaled from zero to $240,000 in annual recurring revenue before entering YC's Winter 2026 batch. Those figures come from founder posts and haven't been independently verified, so take them as aspirational rather than audited.
Real-time insurance checks aren't new in healthcare IT—clearinghouses have offered eligibility verification for years. But weaving that capability into a live AI phone conversation is less common, perhaps because the technical stakes are higher. Get the coverage details wrong mid-call and you've created more problems than you've solved.
The category is crowding up fast. CareCloud, an established EHR and revenue cycle vendor, launched its own stratusAI Desk Agent in December. Tampa General Hospital deployed Hyro's voice AI for call center scheduling last November. Smaller entrants like Clearwave, Powervox, and MedFrontDesk.ai are pitching similar automation. Even UK dental practices are experimenting with voice agents like Boxly.
Pricing, Compliance, and What Comes Next
Patientdesk.ai charges $1,000 per month per location, which includes 1,500 minutes of talk time. Overages run $0.20 per minute. Additional clinic locations cost half the base fee. The pricing calculator on the company's site frames monthly terms as flexible, with a refundable setup fee—an acknowledgment that practices are wary of long contracts and vendor lock-in.
On compliance, the company's privacy policy (last updated early February) states it executes Business Associate Agreements under HIPAA, encrypts data in transit and at rest, and runs on SOC 2-compliant infrastructure. Crucially, the system is non-clinical—it doesn't triage symptoms or offer medical advice, which keeps it clear of trickier regulatory territory.
The three-person team is hiring. They've been invited to exhibit at ADX, a major dental industry event slated for late March. LinkedIn posts from Ozgul frame the ambition as building "the universal AI front desk for healthcare," extending beyond dental into insurance workflows, payments, and care coordination. That's a lot of surface area for a startup barely out of YC.
Whether real-time insurance verification becomes table stakes or remains a differentiator depends on how fast competitors move—and whether clinics trust an AI to quote coverage details accurately enough to collect payment on the spot. For now, Patientdesk.ai is live, taking calls, and growing its clinic count in a market where the phone still rings more often than most practices can answer it. The question isn't whether AI will automate the front desk. It's whose AI gets there first.
